Nonlinear Estimation of Signal Parameters under the Influence of Narrowband Non-Gaussian Noise

The processing of information signal parameters under conditions of inertialess nonlinear transformation of an input mixture of a signal and narrowband non-Gaussian noise is considered. The nonlinear processing of signal parameters is analyzed in the ranges of both small and arbitrary mismatch between the measured signal parameters and their estimates, as well as at a random signal-to-noise ratio at the gauge input. Two methods of optimization of a multichannel discriminator based on the minimum generalized mean-square error criterion and their corresponding structural schemes are obtained. The vector of optimal amplitude characteristics of the nonlinear transformation unit is found. It is shown that the discrimination and fluctuation characteristics of the discriminator are given by the mean value of the error signal and its variance.